Best practices for the publication of Semantic Web data currently place an unacceptably high burden on the end-user, who is supposed to locate and embrace third-party, ontological structures prior to publishing any information. This paper argues for a different publication paradigm where end-users are encouraged to publish potentially incomplete or conflicting information according to their own local context, and where heterogeneous data is consolidated a posteriori through bottom–up, decentralized processes. This approach has two key advantages: first, it greatly simplifies the publication of semantic information by allowing users to contribute purely local data, without any consideration for global schemas or third-party structures. Second, it takes advantage of expressive, semantic links to relate both schemas and instances, creating complex webs of knowledge where cultural or social differences can flourish and be confronted one another.